Investigation of two ferroptosis-related molecular subtypes and biomarkers in the progression of gastric adenocarcinoma

Abstract

This study aimed to investigate ferroptosis-related molecular subtypes and their potential value in gastric adenocarcinoma (GAC). Gene expression and associated clinical phenotype data were downloaded from TCGA database. Unsupervised clustering was performed to identify suitable subtypes, followed by a correlation analysis of clinical factors. Gene set variation analysis, immune infiltration, and differentially expressed genes (DEGs) were investigated between subtypes, followed by functional enrichment for DEGs. Univariate Cox regression analysis was used to screen DEGs to construct a prognostic model. Of the 53 ferroptosis-related genes identified in GAC, two ferroptosis-related molecular subtypes were further examined. These two subtypes showed significant differences in the distribution of histological grades and KEGG pathways. In addition, tumor infiltration of 10 immune cells, including Tregs and natural killer cells, was significantly different between the two subtypes. A total of 660 DEGs were identified between these two subtypes and ultimately 12 genes associated with prognosis, including GPX3, SNCG, DUSP1, and FBN1, were selected by LASSO to construct a prognostic risk model. Survival analysis showed that this model had better prognostic value. Finally, regression analysis showed that the risk score was an independent prognostic indicator of GAC. In summary, two ferroptosis-related molecular subtypes that were associated with the histological grade and tumor immune infiltration were identified. The 12 gene-based risk model had an independent value for predicting GAC prognosis.